mysql查询优化，转自mysql文档

http://tool.oschina.net/uploads/apidocs/mysql-5.5-en/sql-syntax.html#any-in-some-subqueries

Development is ongoing, so no optimization tip is reliable for        the long term. The following list provides some interesting        tricks that you might want to play with:

• Use subquery clauses that affect the number or order of the  rows in the subquery. For example:

  SELECT * FROM t1 WHERE t1.column1 IN
    (SELECT column1 FROM t2 ORDER BY column1);
  SELECT * FROM t1 WHERE t1.column1 IN
    (SELECT DISTINCT column1 FROM t2);
  SELECT * FROM t1 WHERE EXISTS
    (SELECT * FROM t2 LIMIT 1);

• Replace a join with a subquery. For example, try this:

  SELECT DISTINCT column1 FROM t1 WHERE t1.column1 IN (
    SELECT column1 FROM t2);

  Instead of this:

  SELECT DISTINCT t1.column1 FROM t1, t2
    WHERE t1.column1 = t2.column1;

• Some subqueries can be transformed to joins for            compatibility with older versions of MySQL that do not            support subqueries. However, in some cases, converting a            subquery to a join may improve performance. See            Section 13.2.10.11, “Rewriting Subqueries as Joins”.

• Move clauses from outside to inside the subquery. For            example, use this query:

  SELECT * FROM t1
    WHERE s1 IN (SELECT s1 FROM t1 UNION ALL SELECT s1 FROM t2);

  Instead of this query:

  SELECT * FROM t1
    WHERE s1 IN (SELECT s1 FROM t1) OR s1 IN (SELECT s1 FROM t2);

  For another example, use this query:

  SELECT (SELECT column1 + 5 FROM t1) FROM t2;

  Instead of this query:

  SELECT (SELECT column1 FROM t1) + 5 FROM t2;

• Use a row subquery instead of a correlated subquery. For            example, use this query:

  SELECT * FROM t1
    WHERE (column1,column2) IN (SELECT column1,column2 FROM t2);

  Instead of this query:

  SELECT * FROM t1
    WHERE EXISTS (SELECT * FROM t2 WHERE t2.column1=t1.column1
                  AND t2.column2=t1.column2);

• Use NOT (a = ANY (...)) rather than            a <> ALL (...).

• Use x = ANY (table containing            (1,2)) rather than x=1 OR            x=2.

• Use = ANY rather than            EXISTS.

• For uncorrelated subqueries that always return one row,            IN is always slower than            =. For example, use this query:

  SELECT * FROM t1
    WHERE t1.col_name = (SELECT a FROM t2 WHERE b = some_const);

  Instead of this query:

  SELECT * FROM t1
    WHERE t1.col_name IN (SELECT a FROM t2 WHERE b = some_const);

These tricks might cause programs to go faster or slower. Using        MySQL facilities like the        BENCHMARK() function, you can get        an idea about what helps in your own situation. See        Section 12.14, “Information Functions”.

Some optimizations that MySQL itself makes are:

• MySQL executes uncorrelated subqueries only once. Use            EXPLAIN to make sure that a            given subquery really is uncorrelated.

• MySQL rewrites IN,            ALL, ANY, and            SOME subqueries in an attempt to take            advantage of the possibility that the select-list columns in            the subquery are indexed.

• MySQL replaces subqueries of the following form with an            index-lookup function, which            EXPLAIN describes as a            special join type            (unique_subquery or            index_subquery):

  ... IN (SELECT indexed_column FROM single_table ...)

• MySQL enhances expressions of the following form with an            expression involving MIN() or            MAX(), unless            NULL values or empty sets are involved:

  value {ALL|ANY|SOME} {> | < | >= | <=} (uncorrelated subquery)

  For example, this WHERE clause:

  WHERE 5 > ALL (SELECT x FROM t)

  might be treated by the optimizer like this:

  WHERE 5 > (SELECT MAX(x) FROM t)